Helmholtz Graduate School for Polar and Marine Research
ENHANCED SEDIMENT PURIFICATION TECHNIQUES FOR THE <10μm FRACTION OF DIATOM SILICA AND A COMPARISON OF CONTAMINATION ASSESSMENT BEFORE ANALYSING OXYGEN ISOTOPES
B. Chapligin, H. Meyer, U. Hoff, B.Diekmann, A. Eulenburg, H.-W. Hubberten
Alfred Wegener Institute for Polar and Marine Research, Research Unit Potsdam, Telegrafenberg A43, D-14473 Potsdam, Germany bernhard.chapligin@awi.de
Alfred-Wegener-Institute for Polar and Marine Research in the Helmholtz Community
Sediment samples have to go through a long purification process before clean diatom samples can be obtained. This is particularly difficult in the
<10 µm fraction containing clay particles. As the structure of contaminants e.g. organic material, carbonate, mineral particles contains oxygen, it is es- sential to assess the degree of contamination. If high impurities exist a cor- rection factor has to be implemented (Fig. 1). We developed an updated protocol for purifing the <10 µm fraction and compared various methods of assessing the degree of purity to apply the fastest and most appropriate method for future analyses.
Background & Challenge
Contamination „tool kit“ Conclusion
Light Microscope & SEM
Assessment
Light Mic EDS ICP-OES XRF
Assessment
Light Mic EDS ICP-OES XRF
?
?
The analytical possibilities to assess the degree of purity are either by optical identification of contaminants or by chemical analyses determining the chemical composition. Optical methods are represented by light microscope (LM) or scanning electron microscope (SEM). The chemical composition can be either analysed by X-ray Fluorescence (XRF), energy-dispersive X-ray spectroscopy (XRD) under the SEM or inductively coupled plasma optical emission spectrometry (ICP-OES). To compare these methods the biogenic silica standard BFC* was used.
Fig. 1: Contaminants influencing the stable isotope composition of biogenic silica at Lake Baikal.
a) Measured and corrected δ18O values b) Al
2O
3 percentage representing contamination, responsible for the corrected curve in a by Brewer et al (2008)**
Fig. 4: a) Picture under the LM. Contaminants can be identified as particles without regular shape.
The LM is limited, particles smaller than 2 µm cannot be identified.
b) Picure under the SEM. Similar particle detection.
Zooming towards the contaminant supports the identification process. Not limited but slow detec- tion. Convenient for a quick overview or one detailed view only.
The degree of purity under the LM is assessed by chosing random screens and counting the amount of diatoms vs.
contaminants. At least 200, (better 500) items were counted.
Thus, LM is a time taking but precise procedure, as long as the <5 µm fraction (too small for identification) is excluded.
The SEM is not limited by resolution as it is possible to zoom in to every spot not being clearly identified. Again, this is time taking and SEM should only be used as a quick over- view or for detail shots.
If there is more time it is recommended to analyze the sample with ICP-OES taking longer in the preparation phase but giving results with a higher preci- sion. The optical methods don’t provide exact quantified results for the <10 μm fraction but SEM pictures should be used in addition to verify the degree of purity optically by providing detailed view. The EDX is the most recom- mended analysis after this comparison as it needs less than 0.5 mg. In additi- on it is the quickest technique (30/day, n=5) and its precision is good enough for the purpose of assessing degree of purity. Further tests on more biogenic standard material are carried out at the moment to undermine the useful- ness of EDX for contamination assessment.
Fig. 6: EDX analysis of a defined circular area ranging between 200 and 500 µm in diameter. The SiO2 content can be analysed with an average standard deviation of <0.5 % (n=5).
XRF and EDX work with a similar principle of detecting the secondary x-rays after material has been exposed to x-rays. For XRF a high amount of sample material is needed for a reliable analysis (min. 300 mg). EDX is not as precise but needs lesser than 0.5mg as it is opera- ted under the SEM. It provided the fastest results but with a higher mean standard deviation (1σSiO2= 0.5 %). The detection limit for this analysis is comparably high (app. 0.1 %) but still suitable for this pur- The preparation for ICP-OES is time-taking (2 weeks for 30
samples) as full HF, HNO3, HCl digestion has to be performed. As this analysis is normally operated with 100 mg no big advan- tage would have been achieved regarding the amount needed.
Hence, it was attempted to use only 50 mg and 10 mg of sample to see whether this technique is capable of reproducing the known chemical composition. The results were compared to XRF analyses performed in Great Britain on 300 mg BFC*.
b) ICP-OES results with 50 mg and 10 mg are comparable between each other. The final SiO2 content is 97.47 % resp. 97.31 % and thus com- parable to the XRF results.
To analyse the δ18O of diatom material, it is essential to purify the original sediment samples in various physical and chemical preparation steps.
Figure 2 shows the purification process: The sample is freeze-dried to remove water. Then, organic material and cabrbonate are removed by adding H2O2 and HCl for more than 20h on a heating plate at 50°C. The sample is sieved to gain different size fractions (>10 µm, <10µm). The heavy liquid separation (HLS) using sodium-polytungstate (SPT) was repeated 4 times with different solutions of decreasing density (2.4 -2.3 g/cm3). A final acid cleaning is applied to remove micro organics. Several rinsing proce- dures ensured a neutral pH value as well as the complete removal of the SPT solution from the sample.
Preparation steps Effect of preparation steps
Fig. 2: Overview of the preparation steps to gain pure biogenic silica material of 0-10 µm and >10 µm fractions out of a sediment sample.
70.0 75.0 80.0 85.0 90.0 95.0 100.0
1 2 3 4 5 6 7 8
Purification step No.
SiO2 [%]
<10 μm >10 μm
Fig. 3: The development of the SiO2 content throughout the different purification stages (left) was assessed with EDX (see Box „Conta- mination tool kit“). Both fractions (>10 µm, <10 µm) shows a final purity degree of >97 %.The resulting difference in the isotopic com- position is within the analytical error. No correction factor due to contamination has to be applied.
The effect of the different cleaning stages was assessed by using energy dispersive x-ray spectroscopy (see Box „contamination tool kit“) operated under the SEM. The samples were sputtered with Carbon. This is why a quality increase of the sample by removing organic material cannot be observed. The original sediment samples have a SiO2 content of app. 72 %. As more contaminants are (high clay content) is left in the <10 µm fraction, the purity of the >10 µm fraction increases already by sieving. A final purity of >97 % (value shifts below the instrument's error) can be achieved in both fractions. The >10 µm fraction has a purity degree of >97 % alrea- dy after the first heavy liquid separation, where as for the <10 µm fraction the four repetitions of this step are essential.
SEM
XRF & EDX ICP-OES
Sample Freeze-Dry H202; HCl Sieving Heavy Liquid Sep. Acid Cleaning
MS
Organics/Carbonate
Removal of: H2O >2.3 g/cm3 Micro organics
HClO4 + HNO3
<10μm
>10 μm after final acid cl.
after HLS-IV after HLS-III after HLS-II after HLS-I after sieving after H2O2 / HCl dry sample
8 7 6 5 4 3 2 1
Purification step No.
Purification
Method for Contamination assessment
For sediment cores from Lake El'gygytgyn, NE Russia, we found a purification protocol to decrease the non SiO2 fraction to <3 % for the <10 µm as well as for the >10 µm fraction. The major improvement was made by introducing a multiple heavy liquid separation with varying densities. The final acid clea- ning showed no further cleaning effect and can be disregarded in the future.
Method Type min. mass required [mg] Precision time consumption
Light microscope optical count ~ 0.1 0 -
SEM microscope optical count ~ 0.1 0 ++1
EDX analysis chemical analysis < 0.5 + ++
ICP-OES chemical analysis 10 ++ --
XRF chemical analysis 300 +++ -
1just for overview picture, or detailed shot
Fig. 7: Overview table of the various methods for assessing the degree of contamination and their efficiency quality criteria.
* We kindly thank NERC Isotope Geosciences Laboratory (NIGL), British Geological Survey, Keyworth, Nottingham for supplying us with the BFC standard
** published in Brewer TS, Leng MJ, Mackay AW, Lamb AL, Tyler JJ, Marsh NG. 2008. Unravelling contamination signals in biogenic silica oxygen isotope composition: the role of major and trace element geochemistry. Journal of Quaternary Science 23: 321–330 Fig. 5 a) British XRF results on 300 mg BFC
standard material. The loss of ignition is 5.46 %. It is segregated and values are nor- malized to 100%. The final SiO
2 content is 97.22%.**
97.31 97.47 SiO2 to 100%
0.02 bdl
0.04 0.04
0.07 0.08
n.d.
n.d.
0.05 0.06
bdl bdl
0.43 0.45
1.99 1.96
0.09 0.10
dissolved dissolved
10 mg 50mg
100 94.54
0.00 0 P2O5
0.07 0.07 K2O
0.15 0.14 Na2O
0.35 0.33 CaO
0.25 0.24 MgO
0.01 0.01 MnO
0.41 0.39 Fe2O3
1.46 1.38 Al2O3
0.07 0.07 TiO2
97.22 91.91 SiO2
set 100%
XRF by NERC, UK
LOI: 5.46 %
ICP-OES by AWI
P2O5 K2O Na2O CaO MgO MnO Fe2O3 Al2O3 TiO2 SiO2
pose. Again, the analysis (Fig. 6) was compared to the perfor- med XRF of the BFC* standard (Fig. 5).
German Research Foundation This work is funded and supported by:
0.05 0.02 0.05 0.61 0.26 0.02 0.58
0.32 0.02 0.31 97.41 1.34 0.09 0.47
FeO CaO K2O SiO2 Al2O3 MgO Na2O
STDEV AVERAGE
icdp
